Chest Pain in Primary Care: A Systematic Review of Risk Stratification Tools to Rule Out Acute Coronary Syndrome

Search Strategy and Eligibility Criteria

We searched the electronic databases PubMed and Embase for original articles published in English, up to October 9, 2023. A complete overview of the search is presented in Supplemental Appendix 1. Two authors (S.B., A.M.) identified articles for potential inclusion. A third author (R.E.H.) was available to resolve any disagreements. Eligibility criteria were: (1) adults (aged 18 years or more); (2) acute chest pain; (3) enrollment in primary care; (4) available outcome data for ACS, acute myocardial infarction (AMI), or the composite endpoint of major adverse cardiac events (MACE); and (5) use of diagnostic tools. Diagnostic tools were defined as risk stratification tools constructed by a set of condition-specific findings that are applicable in a primary care setting. Exclusion criteria were: studies conducted in ambulance care or emergency department settings or requiring advanced diagnostic testing (eg, coronary angiography, serial biomarker testing); or cardiac monitoring.

Outcomes of Interest

Outcomes of interest were the diagnostic test characteristics of the risk stratification tool: sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and overall discriminative properties (C-statistic). Reference diagnoses of interest were ACS, AMI, and MACE. Major adverse cardiac events are typically defined as a composite of death from any cause, ACS, and urgent coronary interventions. When applicable, we reported the performance of risk stratification tools compared with unaided clinical judgement.

Selection Process

For the selection process, we used an open-source machine learning-aided software (ASReview, Utrecht University) for the initial screening of the titles and abstracts. We followed the principles of the SYstematic review Methodology Blending Active Learning and Snowballing (SYMBALS).^8,9 A more complete description of ASReview’s methods and evidence, including clarification of our stopping criteria can be found in Supplemental Appendix 2. Selected articles were imported into the Covidence systematic review software (Veritas Health Innovation), for full text screening, quality assessment, and data extraction. After reaching the final subset of articles, those reference lists were then screened for possible additional inclusions.

Quality Assessment and Data Extraction

Two researchers (S.B., A.M.) separately extracted data elements from each study. The quality of the included articles was assessed by S.B. and A.M. using the 4 domains of the Quality Assessment of Diagnostic Accuracy Studies-2 tool: patient selection, index test, reference standard, and flow and timing.¹⁰ Risk of bias and applicability concerns were assessed separately. A third author (R.E.H.) independently reviewed the extracted data and quality assessment for accuracy.

Data Synthesis and Analysis

Extracted data on study and patient characteristics, outcome measures, and follow-up information are presented in Table 1. Diagnostic test characteristics and discriminative properties are shown in Table 2.

Search Results

After deduplication, the initial search yielded 1,204 titles and abstracts eligible for inclusion. Following further screening, 53 articles remained, of which 40 were excluded after full text screening. An update of the search resulted in 1 additional article (Figure 1). Backward snowballing yielded no additional inclusions.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Flowchart of systematic review of the literature.

ASR = Active learning for Systematic Reviews.

Study and Patient Characteristics

Table 1 displays the characteristics of the 14 included studies. Nine studies derived or validated a CDR and 5 studies validated troponin assays. Study populations ranged from 115 to 2,192 patients, with follow-up periods from 3 weeks to 6 months. An overview of the components of the included CDRs can be found in Supplemental Table 1. Supplemental Table 2 shows troponin assay characteristics and threshold values.

Quality Assessment

Five of the 14 (35.7%) studies showed a high risk of bias in at least 1 of the quality domains. High risk of bias was most common in the index test and reference standard domains.^15,16,18-20 Applicability concerns were most frequent in patient selection and index test.^11,12,14 A detailed overview of the risk of bias and applicability assessment can be found in Supplemental Appendix 3.

Risk Stratification Tools

A complete overview of diagnostic test characteristics can be found in Table 2.

Strengths and Limitations

This systematic review is an extensive and up-to-date overview of risk stratification tools for acute chest pain in primary care. It will aid GPs in their decision for referral and additional testing. A strength of this review is the inclusion of relatively new strategies using hs-troponin. A limitation is the considerable heterogeneity of the included studies, particularly in patient selection, (a delayed-type) reference standard, and follow-up intervals. Furthermore, several of the included studies were limited by small sample sizes (7 had fewer than 500 patients) and low ACS prevalence. This automatically generates high NPVs and might cause overestimation of the tool’s performance if the sensitivity is not taken into consideration. Finally, we only included articles in English in the review, which might introduce language bias.

Interpretation of Results

Prior international research indicates that a 1% miss rate is considered acceptable for ACS and MACE.^28,29 Additionally, a risk stratification tool also needs to be efficient. Other research found that Dutch GPs would accept 25-50 unnecessary referrals for every patient with ACS.²⁸ Assuming an ACS prevalence of 5%, a specificity of at least 50% is needed. Among the studied CDRs, none met the desired sensitivity and specificity. Only the modified HEART-scores may approach these requirements, but they need further validation. Hence, we cannot recommend currently available CDRs as standalone tools to rule out ACS.

Conventional troponin as a standalone test did not meet the desired diagnostic accuracy, and its use seems to be outdated since the introduction of hs-troponin assays. The studies using hs-troponin measurement show the most promising results with sensitivities up to 100%. Although the study by Andersson reported a lower sensitivity (83%) for their single measurement rule-out strategy, this might be due to using a hs-cTnT threshold below the 99^th percentile.²² In contrast, the study by Johannessen used a much lower threshold below the limit of detection with 100% sensitivity.²³ The European Society of Cardiology guidelines underline the necessity of a very low threshold to rule out ACS by a single hs-troponin measurement and report assay specific thresholds.³⁰

The One-hoUr Troponin in a low-prevalence population of Acute Coronary Syndrome (OUT-ACS) study (published in 2020), prospectively validated the European Society of Cardiology 0/1-hr algorithm, including hs-troponin, to rule out AMI in primary care.^30,31 Although this strategy approaches both desired sensitivity and specificity (98.4% and 79.4%), the algorithm requires serial troponin tests and observation of patients, which is often not possible in primary care settings.

Finding a tool that exceeds the GP’s accuracy might be challenging, since unaided clinical judgment is already reasonably good with a sensitivity 75% to 100% and specificity 18% to 79%.^{1,13-15,19,20} Best diagnostic accuracy might therefore be achieved by combining the GP’s risk assessment with a stratification tool instead of risk stratification tools as standalone tests.

Future Research

Results for risk stratification strategies using hs-troponin are promising, but require further prospective validation, since studies in primary care are limited.^22,23 Although hs-troponin strategies are extensively researched and found to be safe in emergency department settings, results cannot be automatically transferred to primary care due to the lower pre-test probability for ACS.⁵ While the CDRs did not meet the desired test accuracy to rule out ACS, a combination of a CDR with hs-troponin measurement might offer optimal results. Such strategies are currently under research in primary care practices (Pijn op de borst-HELP³²) and out-of-hours primary care (HEART-GP³³).

Variations in health care systems across countries may influence the applicability of risk stratification tools. All but 1 of the included studies were conducted in Europe, where GPs often function as gatekeepers to secondary care. These rule-out strategies, however, could also benefit countries with other primary care structures, such as the United States or countries with large rural regions, where specialized care is scarce and diagnostic options are limited, especially since cardiac monitoring is not required. Such changes, however, require cultural and organizational changes.

Finally, pre-hospital care comprises both primary care and ambulance services and we advocate for standardized care across the entire continuum. Future research should focus on risk stratification tools that are applicable in the entire pre-hospital care chain, and preferably in line with hospital care. The latter is especially important for strategies using troponin since different assays and thresholds hamper communication between health care providers.